Data Mining

Subject: The Economist article on data mining

http://www.economist.com/world/international/displaystory.cfm?story_id=12295455 
 >

Know-alls

Sep 25th 2008
> From The Economist print edition

...

Last month, after a briefing by the Department of Justice about a  
secret data-mining plan for the FBI, a group of American lawmakers  
wrote to Mr Mukasey complaining that the plan would allow the FBI to  
spy on Americans "without any basis for suspicion". The proposed  
project could be made public in coming weeks.

No similar pan-European data-mining programme is operating, at least  
to public knowledge. Yet under an agreement signed in July last year  
airlines flying from the European Union to America have had to provide  
the authorities there with reservations data, as well as information  
obtained by airport-security screeners. This can include passengers'  
race, religion, occupation, relatives, hotel reservations and credit  
card details. Internet service providers and telecoms firms in the EU  
must now keep for up to two years, though not automatically hand over,  
data on websites visited and phone calls made and received (but not  
the content of conversations).

...

Spies are increasingly snooping on private internet use. Katharina von  
Knop, a data-mining expert at the University of German Federal Armed  
Forces in Munich, says many systems remotely analyse the content of  
web pages people visit. A man who has travelled to, say, Peshawar, a  
stronghold of Islamist extremism in Pakistan, is considered more  
dangerous if he also reads the blog of an extremist Muslim cleric. If  
the cleric lives in Peshawar, the man's suspicion score rises further.  
Data-mining software develops profiles by taking into account all web  
pages visited by a computer user; if a suspect visits a stamp- 
collecting website, the suspicion score is lowered.

Such profiling increasingly relies on "sentiment analysis". Hsinchun  
Chen, head of the Artificial Intelligence Lab at the University of  
Arizona says this technique, which he performs for American and  
international intelligence agencies, is an emerging and booming field.  
The goal is to identify changes in the behaviour and language of  
internet users that could indicate that angry young men are becoming  
potential suicide-bombers. For example, a person who exhibits  
curiosity by visiting many Islamist websites and asking numerous  
questions in online forums might be flagged by sentiment-analysis  
software if he shows signs of resentment and eventually turns to  
"radicalising" others by, say, justifying violence and providing links  
to militant videos. Mr Chen says intelligence agencies in the United  
States, Canada, China, Germany, Israel, Singapore and Taiwan are  
customers for this technique.


Does it work?

Donald Tighe, vice-president for public affairs at In-Q-Tel, a non- 
profit investment outfit that helps the CIA stay abreast of advances  
in computing, says that data mining is now so powerful it has become  
"essential to our national security". But campaigners for privacy have  
many worries. One fear, prevalent in Britain after incidents in which  
officials lost huge quantities of confidential personal information,  
is that the state may be even more careless with data than private  
firms are. Another is that innocents are flagged for further  
investigation or added to "watch-lists" that may impede air travel,  
banking and gaining jobs in places where radioactive materials are  
used, such as hospitals. The American Civil Liberties Union (ACLU), a  
lobby, says the list maintained by the Terrorist Screening Centre at  
the FBI now has more than 900,000 names, with 20,000 more every month.  
Being removed is tricky.

Data-mining may be bad for national security as well as for civil  
liberties. The software is often modelled on the fraud-detection  
applications used by financial institutions. But terrorism is much  
rarer. So spotting conditions that may precede attacks is harder. Mike  
German, a former FBI agent who now advises the ACLU, says intelligence  
agencies too readily believe in the "snake oil" of total information  
awareness, which drains effort from more useful activities such as  
using informers and infiltrators.